CC BY
8the LOGOS-Aero module of the LOGOS software package. Performance of LOGOS software components
employing Lagrangian and Eulerian multi-phase flow models is demonstrated by test simulations of some
NACA problems [4] intended for verification of the Lewice software package.
References
1. M. A. Pogosyan, E. P. Savelievskikh, R. M. Shagaliev, A. S. Kozelkov, D. Yu. Strelets, A. A. Ryabov, A. V. Kornev, Yu.
N. Deryugin, V. F. Spiridonov, K. V. Tsiberev. Application of Russian supercomputer technologies to develop the advanced
models of aviation technology // Voprosy atomnoy nauki i tekhniki. Ser. Mathematical Modeling of Physical Processes
2013. Iss. 2. P. 3-18. [In Russian].
2. Kozelkov A.S., Zhuchkov R.N., Utkina A.A., Volodchenkova K.B. Simulation of turbulent flows with higher-order
schemes on hybrid-structure grids // J. VANT, Ser. Mathematical Modeling of Physical Processes, 2014, issue 3, p. 18-31.
[In Russian].
3. Betelin V.B., Shagaliev R.M., Aksenov S.V., Belyakov I.M., Deryuguin Yu.N., Kozelkov A.S., Korchazhkin D.A., Nikitin
V.F., Sarazov A.V., Zelenskiy D.K. Mathematical simulation of hydrogen�oxygen combustion in rocket engines using
LOGOS code // Acta Astronautica 2014, v. 96, p.53�64.
4. Wright B.W., Rutkowski A. Validation Results for LEWICE 2.0, NASA/CR�1999-208690, 1999.
Comparison of MKL matrix multiplication routines for one practical example
V. S. Gladkikh, Y. L. Gurieva
Institute of Computational Mathematics and Mathematical Geophysics SB RAS
Email: gladvs_ru@mail.com
DOI 10.24412/cl-35065-2021-1-01-70
Nowadays, math libraries (MKL [3] and Netlib BLAS [1]) are used to get the best performance of applica-
tion. Extensive libraries� functionality often allows applied program to be implemented via various library pro-
cedures that have different levels of optimization. As a result decision about which routine should be used is a
non-trivial task. A roof-line model [2] can help to identify some weak points of the software and prepare re-
quired experiments that identify the optimal library procedure. Given one specific practical example, it was
shown that MKL BLAS gemm routine preferable over the similar MKL BLAS gemv procedure for the target set
of the input data.
References
1. Dongarra J.J. [� ��.]. A set of level 3 basic linear algebra subprograms // ACM Transactions on Mathematical
Software. 1990. No. 1 (16). C. 1�17.
2. Ofenbeck G. [� ��.]. Applying the roofline model // ISPASS 2014 - IEEE International Symposium on Performance
Analysis of Systems and Software. 2014. C. 76�85.
3. Intel Intel Math Kernal Library(MKL) [web]. URL: http://software.intel.com/en-us/articles/intel-math-kernel-
library-documentation.
LOGOS software package. Heat-transfer problem solving method with the account for ablation process
V. A. Glazunov, Yu. D. Seryakov, R. A. Trishin
FSUE �Russian Federal Nuclear Center � All-Russian Research Institute of Experimental Physics�, Sarov, Nizhny
Novgorod Region
�mail: staff@vniief.ru
DOI 10.24412/cl-35065-2021-1-01-71
An approach to simulate 3D heat-transfer problem with the account for the surface ablation process is re-
alized in the LOGOS Thermal Analysis product [1]. The urgency of the work comes from the need for the ade-
quate thickness definition of the low-conductivity coating of the flying vehicle during its operation.
Stefan exterior problem is solved for the computation of the boundary nodes displacement in the grid
model. Grid internal nodes motion is computed using an elastic smoothing method. Quality preservation of
the grid model during the computation is the application condition for the approach. Three cell quality criteria
are considered, as well as additional possibilities to preserve grid topology.
A step-by-step computational algorithm is proposed using the example of the heat-transfer problem solu-
tion with the account for the design shape changing; it includes the computation until the stopping criterion is
met, remeshing and proceeding with the computation.
References
1. Deryugin Yu.N., Zelensky D.K., Glazunov V.A. et al. LOGOS multifunctional software package: physical and
mathematical computational models for aero- and hydrodynamics and heat transfer: Preprint. RFNC-VNIIEF. 111-2013.
Sarov: RFNC-VNIIEF, 2013.
Some approaches to the creation of supercomputer technologies for solving compute-intensive problem
B. �. Glinskiy1, D. V. Wins1, Y. A. Zagorulko2, G. B. Zagorulko2, I. M. Kulikov1, A. F. Sapetina1, P. A. Titov1,
I. G. Chernykh1
1Institute of Computational Mathematics and Mathematical Geophysics SB RAS
2A. P. Ershov Institute of Informatics System SB RAS
Email: gbm@sscc.ru
DOI 10.24412/cl-35065-2021-1-01-72
The issues of using supercomputer technologies developed by the authors of the article to solve compute-
intensive problems are discussed. The developed technology of creating algorithmic and software for super-
computers contains three related stages: co-design, by which we understand the adaptation of the problem
statement, the mathematical method, the computational algorithm to the parallel architecture of the super-
computer at all stages of the problem solution; study of the scalability of computational algorithms for the
most promising supercomputers based on simulation modeling; evaluation of the energy efficiency of algo-
rithms for various implementations on a given supercomputer architecture [1]. It is proposed to further devel-
op the proposed approach with the use of intellectual support for solving computationally complex problems
using the ontology of computational methods and algorithms for solving the problem, the ontology of compu-
tational heterogeneous architectures and decision rules [2]. For clarity, illustrations are presented that sche-
matically display the various components of the process of solving a geophysical problem, from stating to im-
plementation on a supercomputer, and also how those are interconnected [3]. An example of field observation
processing for one of the areas of Western Siberia using the developed system is presented [4].
The work was carried out within the framework of the budget project of the ICMMG SB RAS 0251-2021-0005 (sec-
tion 3), as well as with the support of the RFBR grants No 19-07-00085 (sections 2, 3).
References
1. Glinskiy, B., Kulikov, I., Chernykh, I., Snytnikov, A., Sapetina, A., Weins, D. The integrated approach to solving large-
size physical problems on supercomputers (2017) Communications in Computer and Information Science, 793, pp. 278-
289. DOI: 10.1007/978-3-319-71255-0_22.
2. B. Glinskiy, Y. Zagorulko, G. Zagorulko, I. Kulikov, A. Sapetina. The Creation of Intelligent Support Methods for
Solving Mathematical Physics Problems on Supercomputers. Russian Supercomputing Days 2019, Springer International
Publishing 2019, 427-438, DOI 10.1007/978-3-030-36592-9_35.
